Job description
Computer Vision Engineer
Fast Code AI · Bengaluru · Full-time · Software Development
Responsibilities
- Design and implement 2D/3D computer vision algorithms and downstream processing
- Build ML pipelines for data handling, auto-labelling, training, and validation
- Train segmentation and detection models on modern backbones (e.g., DINO, Mask2Former, Point Transformer)
- Develop camera–LiDAR fusion approaches for perception
- Write robust, efficient, well-tested production code
- Analyze metrics (mIoU, per-class IoU) and optimize models
- Solve sensor calibration and alignment as optimization problems, and set measurable accuracy benchmarks for the results
- Create tooling to inspect projection quality, reprojection residuals and cross-sensor label consistency
- Streamline data processing through automated pipelines and CI/CD, and take part in code and design reviews
- BTech/MTech with proven experience in computer vision and ML
- Strong C++ and Python (PyTorch/TensorFlow)
- Deep understanding of object detection, semantic/panoptic segmentation, keypoint detection
- Knowledge of traditional CV (SIFT, camera calibration, fundamental matrix)
- 2 year experience with production ML pipelines, from dataset collection to deployment
- Proficiency with debugging/profiling tools (Nvidia Nsight, etc.)
- ADAS / autonomous driving perception experience, incl. 3D point clouds
- Publications in top-tier CV/ML venues
- Track record of building algorithms from scratch
What We Value
- Hands-on data analysis and optimization
- Fast iteration in early-stage development
- Clear technical communication and documentation
Keywords
TransformersData AnalyticsContinuous Integration (CI)CalibrationDownstream Oil and GasOptimizationProjectionDownstream ProcessingAlgorithmsTraining
Interested in this role?